Nobody can agree how much savings Americans have left

The Federal Reserve Bank of San Francisco reckons that about $500 billion of the $2.1 trillion in excess pandemic savings is left, and that it will be spent by late summer. Other research teams on Wall Street say households still have more than $1 trillion. The simple question of how much cash Americans have left has become the main macroeconomic argument of 2023, and the answer depends on which model an economist picks. The gap between estimates matters, because it decides how long consumers can keep spending.

The disagreement comes from how "excess savings" are calculated. Nobody can see them on a bank statement. They are a residual: the difference between what households actually saved and what they would have saved had the pandemic never happened.

The pile was built by policy as much as by lockdown. Three rounds of stimulus payments, worth $1,200, $600 and $1,400 a person, together with enhanced unemployment benefits and an expanded child tax credit, pushed disposable income above its trend just as restaurants, travel and entertainment were closed. The personal saving rate hit a record in April 2020. Lower-income households spent much of their share on goods as soon as it arrived, which is one reason the money ran out sooner at the bottom.

It depends on the baseline

Assume that households would have kept saving at the pre-pandemic rate of 7.5 per cent seen in 2018–2019, and actual savings today look very low. On that basis the excess is almost used up.

Use a longer historical trend, or allow for faster growth in nominal incomes, and a sizeable surplus remains. Aggregate models also say nothing about who holds the money. Most of the remaining liquid cash sits with households in the top fifth of the income distribution, who treat it as wealth to keep, not money to spend.

The data are also subject to revision. Saving is calculated as income minus spending, both of which are estimated and later revised, so a small change in either can move the cumulative total by tens of billions of dollars. An estimate built on two years of monthly residuals carries the errors of every month in it.

Inflation shrinks the pile

Every estimate is in nominal dollars, and that flatters them all. Consumer prices have risen by more than a tenth since the savings peaked in 2021, so a dollar saved during lockdown buys noticeably less today. Measured in what it can purchase, the remaining cushion is smaller than any of the headline figures suggest.

Higher interest rates pull the other way for the households that still hold cash. Money market funds and certificates of deposit now pay close to 5 per cent, which gives affluent savers a reason to leave their surplus where it is rather than spend it. The remaining savings may therefore behave less like a spending reserve and more like ordinary wealth.

The bottom has run dry

For the bottom sixty per cent of earners, the pandemic cushion is gone. Credit card balances have climbed in its place, and auto loan delinquencies are in double digits.

The New York Fed's latest household debt report, published this week, showed total household debt passing $17 trillion in the first quarter, with credit card balances close to $1 trillion. Card balances usually fall in the first quarter as households pay off holiday spending; this year they barely did. Borrowing at card rates above 20 per cent is what households do when savings are gone and incomes do not cover the bills.

What the retailers see

The companies selling to these households are reporting the split directly. Target said this week that its shoppers were cutting back on discretionary items and focusing on essentials. Walmart, by contrast, raised its outlook for the year, helped in part by better-off shoppers trading down to cheaper stores. Both patterns fit a consumer base in which the top has money and the rest are economising.

That divergence matters for which businesses suffer when the labour market turns. Discount chains and lenders to lower-income borrowers face customers with no buffer, while premium brands and travel companies can lean on households that still hold cash.

The case for resilience

The strongest counterargument is that the labour market, not savings, is what really drives spending. Unemployment was 3.4 per cent in April, and as long as people have jobs and pay rises, consumption can continue even with a thinner cushion. Most homeowners also locked in low fixed mortgage rates during the pandemic, which shields them from the rate rises that would otherwise squeeze their budgets.

That holds while hiring stays strong. It makes spending more sensitive to jobs, not less, because households without savings cannot smooth their spending through a spell of unemployment. The Supreme Court's ruling on the administration's student-loan forgiveness plan, expected by the end of June, and the eventual resumption of repayments will add another test for the same households.

What would settle it

The monthly personal income and outlays report, due at the end of the month, will show whether the saving rate is still falling. Payroll figures in early June will show whether hiring is slowing. Delinquency rates on cards and car loans are the most direct gauge of which households have run out of room.

Using the aggregate figure to predict how well consumers will hold up is statistical self-deception. Economists cannot agree on the number because the number is a modelling construct. For most households the cash buffer has already been spent, so consumer spending will feel any weakening in the labour market later this year almost immediately.